Network Management

Hi all,
I think I have a problem with a multi chassis lag created on a pair of 8100 in vsx mode.
Let me explain: 

I have a pair of 8100s connected in vsx that I want to connect to an HPE 7503

I create my lag on 8100:

vsxa:
interface lag 100 multi-chassis
description ssfc-irf-lag
no shutdown
no routing
vlan trunk native 1
vlan trunk allowed all
lacp mode active
interface 1/1/1
description ssfc-irf-lag
no shutdown
mtu 9100
lag 100
vsx b:
interface lag 100 multi-chassis
description ssfc-irf-lag
no shutdown
no routing
vlan trunk native 1
vlan trunk allowed all
lacp mode active
interface 1/1/1
description ssfc-irf-lag
no shutdown
mtu 9100
lag 100
hpe 7500:
interface Bridge-Aggregation200
description vsx
port link-type trunk
port trunk permit vlan all
link-aggregation mode dynamic
dhcp snooping trust
interface Ten-GigabitEthernet1/0/0/5
port link-mode bridge
description vsx
port link-type trunk
port trunk permit vlan all
port link-aggregation group 200
interface Ten-GigabitEthernet1/1/0/5
port link-mode bridge
description vsx
port link-type trunk
port trunk permit vlan all
port link-aggregation group 200

here's what the various devices see:

7503
display interface Bridge-Aggregation 200 brief
Brief information on interfaces in bridge mode:
Link: ADM - administratively down; Stby - standby
Speed: (a) - auto
Duplex: (a)/A - auto; H - half; F - full
Type: A - access; T - trunk; H - hybrid
Interface Link Speed Duplex Type PVID Description
BAGG200 UP 20G(a) F(a) T 1 vsx
vsxA:
show interface brief
lag100 1 trunk -- yes up -- 10000 ssfc-irf-lag
vsxB:
show interface brief
lag100 1 trunk -- yes up -- 10000 ssfc-irf-lag

is everything normal? I'm a bit confused.

tnx

Pietro

As example, let me consider a LAG between a VSX Cluster (VSX LAG with LACP to the VSF Virtual Chassis) and a VSF - a VSF is a "Virtual Chassis" made of two Aruba modular switches - (simple LAG with LACP to VSX Cluster) [*]...then, on both VSX Primary and VSX Secondary, you would have a VSX LAG (MC-LAG) definition like that:

interface lag 254 multi-chassis
    description Aruba-VSF-Trk1
    no shutdown
    no routing
    vlan trunk native <PVID> <------------ in case you need a tagged native then add the "tag" at the end
    vlan trunk allowed <PVID> + <other-IDs-for-allowing-tagged-VLAN>
    lacp mode active
    loop-protect
    loop-protect vlan <PVID> + <other-IDs-for-allowing-tagged-VLAN>
    exit

and then, on VSX Primary:

interface 1/1/29
    description Aruba-VSF-Trk1-1/A1
    no shutdown
    mtu 9198
    flow-control rx
    lag 254
interface 1/1/30
    description Aruba-VSF-Trk1-2/A1
    no shutdown
    mtu 9198
    flow-control rx
    lag 254

On VSX Secondary:

interface 1/1/29
    description Aruba-VSF-Trk1-1/B1
    no shutdown
    mtu 9198
    flow-control rx
    lag 254
interface 1/1/30
    description Aruba-VSF-Trk1-2/B1
    no shutdown
    mtu 9198
    flow-control rx
    lag 254

The output of show lacp interfaces multi-chassis on the VSX Primary would then be (redacted):

State abbreviations :
A - Active        P - Passive      F - Aggregable I - Individual
S - Short-timeout L - Long-timeout N - InSync     O - OutofSync
C - Collecting    D - Distributing 
X - State m/c expired              E - Default neighbor state

Actor details of all interfaces:
----------------------------------------------------------------------------------
Intf       Aggregate  Port    Port     State   System-ID         System   Aggr    
           name       id      Priority                           Priority Key     
----------------------------------------------------------------------------------
1/1/29     lag254(mc) 29      1        ALFNCD  02:02:00:xx:xx:xx 65534    254     
1/1/30     lag254(mc) 30      1        ALFNCD  02:02:00:xx:xx:xx 65534    254     

Partner details of all interfaces:
----------------------------------------------------------------------------------
Intf       Aggregate  Partner Port     State   System-ID         System   Aggr    
           name       Port-id Priority                           Priority Key     
----------------------------------------------------------------------------------
1/1/29     lag254(mc) NNN     1        ALFNCD  f4:03:43:yy:yy:yy 4159     963     
1/1/30     lag254(mc) MMM     1        ALFNCD  f4:03:43:yy:yy:yy 4159     963     

Remote Actor details of all interfaces:
----------------------------------------------------------------------------------
Intf       Aggregate  Port    Port     State   System-ID         System   Aggr    
           name       id      Priority                           Priority Key     
----------------------------------------------------------------------------------
1/1/30     lag254(mc) 1030    1        ALFNCD  02:02:00:xx:xx:xx 65534    254     
1/1/29     lag254(mc) 1029    1        ALFNCD  02:02:00:xx:xx:xx 65534    254     

Remote Partner details of all interfaces:
----------------------------------------------------------------------------------
Intf       Aggregate  Partner Port     State   System-ID         System   Aggr    
           name       Port-id Priority                           Priority Key     
----------------------------------------------------------------------------------
1/1/30     lag254(mc) PPP     1        ALFNCD  f4:03:43:yy:yy:yy 4159     963     
1/1/29     lag254(mc) QQQ     1        ALFNCD  f4:03:43:yy:yy:yy 4159     963

Please note the ALFNCD status on lag254 as seen on Actor and Remote Actor (VSX members) and they related Partner and Remote Partner peers.

[*] Same approach if you consider your scenario where a VSX Cluster connects to a single modular Chassis represented by an HPE 7503 (I imply that you aren't in a IRF scarnio, isn'it? in any case - also with IRF - nothing really should change on the approach reported above).

Cheers.

I believe we see the speed of the aggregated interface that each VSX only owns, not the grand total seen by the VSX Cluster. I have VSX LAGs with 3+3 (3x10 + 3x10) and I see 30Gbps not 60Gbps but I believe it is normal in the case of Multi-Chassis LAGS. Now I've no time but I will go deep to see if, somewhere, the ArubaOS-CX reports the VSX LAG bandwidth grand total.

Hi @DP-7f7ca3, 

I have the same question. Why doesn't the 'show interface brief' or 'show interface' command display the total speed value? Did you manage to get an answer?

Hello @rodrigo_santos, unfortunately I never found a workaround or a reason. But I was referring to the fact that a VSX LAG logical interface (a multi-chassis LAG interface) simply shows the grand total of speeds of its memeber interfaces (e.g. 3x10 on a VSX member) exactly and only on the particular VSX member you're considering (when you run the command), that's it (that is because the dual plane personality of the VSX clustering). I don't believe HPE changed anything to let (any related) command to show the total aggregated speed considering the same VSX LAG as an entity distributed on both the VSX members (e.g. 3x10 + 3x10 = 60 for both VSX members).

-------------------------------------------